public inbox for glibc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "tatannen at wisc dot edu" <sourceware-bugzilla@sourceware.org>
To: glibc-bugs@sources.redhat.com
Subject: [Bug libc/13601] New: Fail to load dynamic libs when entire ELF header cannot be read in one syscall
Date: Wed, 18 Jan 2012 04:15:00 -0000	[thread overview]
Message-ID: <bug-13601-131@http.sourceware.org/bugzilla/> (raw)

http://sourceware.org/bugzilla/show_bug.cgi?id=13601

             Bug #: 13601
           Summary: Fail to load dynamic libs when entire ELF header
                    cannot be read in one syscall
           Product: glibc
           Version: 2.14
            Status: NEW
          Severity: normal
          Priority: P2
         Component: libc
        AssignedTo: drepper.fsp@gmail.com
        ReportedBy: tatannen@wisc.edu
    Classification: Unclassified


As stated in the man pages, the read() system call may read/return less bytes
than requested, for a variety of well-documented reasons.
In elf/dl-load.c, function open_verify() attempts to read the entire elf header
(over 800 bytes on x86_84) in one invocation of the read system call, and if
all bytes are not read, the shared library will not load with a "file too
short" error. This could cause failures, for instance, loading over networked
file systems configured with a small (less than 800byte) MTU size.

In elf/dl-load.c the problem is here:
1714:      size_t maplength;
1715:
1716:      /* We successfully openened the file.  Now verify it is a file
1717:     we can use.  */
1718:      __set_errno (0);
1719:      fbp->len = __libc_read (fd, fbp->buf, sizeof (fbp->buf));
1720:
1721:      /* This is where the ELF header is loaded.  */
1722:      assert (sizeof (fbp->buf) > sizeof (ElfW(Ehdr)));
1723:      ehdr = (ElfW(Ehdr) *) fbp->buf;
1724:
1725:      /* Now run the tests.  */
1726:      if (__builtin_expect (fbp->len < (ssize_t) sizeof (ElfW(Ehdr)), 0))
1727:    {
1728:      errval = errno;

On line 1719, if the read returns less than sizeof (ElfW(Ehdr)) as tested on
line 1726, the library load will fail.

The call to libc_read on line 1719 should be repeated while libc_read
returns a value greater than zero.  If you desire me to write/submit a patch,
I'd be happy to help out, just ask.

Thank you!

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


             reply	other threads:[~2012-01-18  4:15 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-01-18  4:15 tatannen at wisc dot edu [this message]
2012-02-21  2:34 ` [Bug ld.so|libdl/13601] " jsm28 at gcc dot gnu.org
2012-10-20  3:31 ` [Bug dynamic-link/13601] " siddhesh at redhat dot com
2014-06-27 11:11 ` fweimer at redhat dot com

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=bug-13601-131@http.sourceware.org/bugzilla/ \
    --to=sourceware-bugzilla@sourceware.org \
    --cc=glibc-bugs@sources.redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).